"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "emotion = 'happy'\n",
+ "path = np.array(df['speech'][df['label']==emotion])[0]\n",
+ "data, sampling_rate = librosa.load(path)\n",
+ "waveplot(data, sampling_rate, emotion)\n",
+ "spectogram(data, sampling_rate, emotion)\n",
+ "Audio(path)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 17,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def extract_mfcc(filename):\n",
+ " y, sr = librosa.load(filename, duration=3, offset=0.5)\n",
+ " mfcc = np.mean(librosa.feature.mfcc(y=y, sr=sr, n_mfcc=40).T, axis=0)\n",
+ " return mfcc"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 18,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "array([-4.0976691e+02, 8.7749123e+01, 2.2847012e+01, 4.0431747e+01,\n",
+ " 9.7905273e+00, 3.7332246e+00, -2.7344184e+00, -8.9495554e+00,\n",
+ " 3.5624874e+00, -1.2005082e+01, -3.6908846e+00, 1.8095992e+00,\n",
+ " -5.3735299e+00, 6.9867735e+00, -6.2379251e+00, 1.0952128e+00,\n",
+ " 7.6578659e-01, -2.4418123e+00, -7.9200463e+00, -5.9605379e+00,\n",
+ " -7.1019883e+00, -7.7682948e+00, -3.6830564e+00, 4.1091146e+00,\n",
+ " -3.0087498e-01, 5.9679031e+00, 2.0398079e-01, -6.6093755e-01,\n",
+ " -2.6439133e+00, -2.0392258e-01, 3.6369934e+00, 9.6661215e+00,\n",
+ " 8.0118532e+00, 7.1307998e+00, 2.2150338e+00, 1.8789151e+00,\n",
+ " 2.0025277e+00, 1.9759331e+00, 4.3048563e+00, 3.0489488e+00],\n",
+ " dtype=float32)"
+ ]
+ },
+ "execution_count": 18,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "extract_mfcc(df['speech'][0])"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "X_mfcc = df['speech'].apply(lambda x: extract_mfcc(x))"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "0 [-409.7669, 87.74912, 22.847012, 40.431747, 9....\n",
+ "1 [-438.59506, 66.7165, 23.917074, 38.291676, 6....\n",
+ "2 [-432.72467, 107.85921, 25.550182, 30.242085, ...\n",
+ "3 [-345.25906, 94.46191, 11.120116, 24.519775, -...\n",
+ "4 [-395.94043, 111.45156, 27.71482, 14.070998, -...\n",
+ " ... \n",
+ "5595 [-374.46442, 88.345566, -8.879825, -18.313759,...\n",
+ "5596 [-437.882, 86.159096, -5.292689, -23.754158, -...\n",
+ "5597 [-428.6321, 100.859566, -27.002722, -40.025597...\n",
+ "5598 [-434.54526, 82.24351, 17.994448, -10.922421, ...\n",
+ "5599 [-468.8798, 63.885868, -2.7314255, -7.891333, ...\n",
+ "Name: speech, Length: 5600, dtype: object"
+ ]
+ },
+ "execution_count": 20,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "X_mfcc"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 21,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "(5600, 40)"
+ ]
+ },
+ "execution_count": 21,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "X = [x for x in X_mfcc]\n",
+ "X = np.array(X)\n",
+ "X.shape"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "(5600, 40, 1)"
+ ]
+ },
+ "execution_count": 22,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "## input split\n",
+ "X = np.expand_dims(X, -1)\n",
+ "X.shape"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from sklearn.preprocessing import OneHotEncoder\n",
+ "enc = OneHotEncoder()\n",
+ "y = enc.fit_transform(df[['label']])"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "y = y.toarray()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 25,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "(5600, 7)"
+ ]
+ },
+ "execution_count": 25,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "y.shape"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 26,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "2024-07-17 22:08:57.690085: I external/local_xla/xla/tsl/cuda/cudart_stub.cc:32] Could not find cuda drivers on your machine, GPU will not be used.\n",
+ "2024-07-17 22:08:58.721701: I external/local_xla/xla/tsl/cuda/cudart_stub.cc:32] Could not find cuda drivers on your machine, GPU will not be used.\n",
+ "2024-07-17 22:08:59.469766: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:485] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n",
+ "2024-07-17 22:09:00.003130: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:8454] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n",
+ "2024-07-17 22:09:00.129667: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1452] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n",
+ "2024-07-17 22:09:01.247941: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
+ "To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
+ "2024-07-17 22:09:18.319439: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n"
+ ]
+ },
+ {
+ "data": {
+ "text/html": [
+ "Model: \"sequential\"\n",
+ "
\n"
+ ],
+ "text/plain": [
+ "\u001b[1mModel: \"sequential\"\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
+ "┃ Layer (type) ┃ Output Shape ┃ Param # ┃\n",
+ "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
+ "│ lstm (LSTM) │ (None, 256) │ 264,192 │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dropout (Dropout) │ (None, 256) │ 0 │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense (Dense) │ (None, 128) │ 32,896 │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dropout_1 (Dropout) │ (None, 128) │ 0 │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_1 (Dense) │ (None, 64) │ 8,256 │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dropout_2 (Dropout) │ (None, 64) │ 0 │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_2 (Dense) │ (None, 7) │ 455 │\n",
+ "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
+ "
\n"
+ ],
+ "text/plain": [
+ "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
+ "┃\u001b[1m \u001b[0m\u001b[1mLayer (type) \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
+ "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
+ "│ lstm (\u001b[38;5;33mLSTM\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m) │ \u001b[38;5;34m264,192\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dropout (\u001b[38;5;33mDropout\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m) │ \u001b[38;5;34m0\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m128\u001b[0m) │ \u001b[38;5;34m32,896\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dropout_1 (\u001b[38;5;33mDropout\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m128\u001b[0m) │ \u001b[38;5;34m0\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_1 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m64\u001b[0m) │ \u001b[38;5;34m8,256\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dropout_2 (\u001b[38;5;33mDropout\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m64\u001b[0m) │ \u001b[38;5;34m0\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_2 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m7\u001b[0m) │ \u001b[38;5;34m455\u001b[0m │\n",
+ "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ " Total params: 305,799 (1.17 MB)\n",
+ "
\n"
+ ],
+ "text/plain": [
+ "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m305,799\u001b[0m (1.17 MB)\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ " Trainable params: 305,799 (1.17 MB)\n",
+ "
\n"
+ ],
+ "text/plain": [
+ "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m305,799\u001b[0m (1.17 MB)\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ " Non-trainable params: 0 (0.00 B)\n",
+ "
\n"
+ ],
+ "text/plain": [
+ "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "from keras.models import Sequential\n",
+ "from keras.layers import Dense, LSTM, Dropout\n",
+ "\n",
+ "model = Sequential([\n",
+ " LSTM(256, return_sequences=False, input_shape=(40,1)),\n",
+ " Dropout(0.2),\n",
+ " Dense(128, activation='relu'),\n",
+ " Dropout(0.2),\n",
+ " Dense(64, activation='relu'),\n",
+ " Dropout(0.2),\n",
+ " Dense(7, activation='softmax')\n",
+ "])\n",
+ "\n",
+ "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
+ "model.summary()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 27,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Epoch 1/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m25s\u001b[0m 278ms/step - accuracy: 0.5024 - loss: 1.3013 - val_accuracy: 0.8929 - val_loss: 0.2595\n",
+ "Epoch 2/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m19s\u001b[0m 262ms/step - accuracy: 0.9158 - loss: 0.2639 - val_accuracy: 0.9321 - val_loss: 0.1905\n",
+ "Epoch 3/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 236ms/step - accuracy: 0.9462 - loss: 0.1857 - val_accuracy: 0.9705 - val_loss: 0.0838\n",
+ "Epoch 4/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m23s\u001b[0m 268ms/step - accuracy: 0.9632 - loss: 0.1241 - val_accuracy: 0.9205 - val_loss: 0.2495\n",
+ "Epoch 5/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 264ms/step - accuracy: 0.9632 - loss: 0.1145 - val_accuracy: 0.9464 - val_loss: 0.1530\n",
+ "Epoch 6/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 265ms/step - accuracy: 0.9652 - loss: 0.1152 - val_accuracy: 0.9875 - val_loss: 0.0341\n",
+ "Epoch 7/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 262ms/step - accuracy: 0.9761 - loss: 0.0700 - val_accuracy: 0.9821 - val_loss: 0.0540\n",
+ "Epoch 8/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 266ms/step - accuracy: 0.9772 - loss: 0.0877 - val_accuracy: 0.9777 - val_loss: 0.0691\n",
+ "Epoch 9/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 269ms/step - accuracy: 0.9830 - loss: 0.0593 - val_accuracy: 0.9741 - val_loss: 0.0776\n",
+ "Epoch 10/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 261ms/step - accuracy: 0.9851 - loss: 0.0484 - val_accuracy: 0.9884 - val_loss: 0.0240\n",
+ "Epoch 11/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 265ms/step - accuracy: 0.9881 - loss: 0.0393 - val_accuracy: 0.9241 - val_loss: 0.2182\n",
+ "Epoch 12/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 268ms/step - accuracy: 0.9638 - loss: 0.1245 - val_accuracy: 0.9786 - val_loss: 0.0641\n",
+ "Epoch 13/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 265ms/step - accuracy: 0.9802 - loss: 0.0691 - val_accuracy: 0.9857 - val_loss: 0.0379\n",
+ "Epoch 14/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 265ms/step - accuracy: 0.9906 - loss: 0.0299 - val_accuracy: 0.9929 - val_loss: 0.0284\n",
+ "Epoch 15/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 231ms/step - accuracy: 0.9841 - loss: 0.0436 - val_accuracy: 0.9937 - val_loss: 0.0161\n",
+ "Epoch 16/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 242ms/step - accuracy: 0.9942 - loss: 0.0149 - val_accuracy: 0.9929 - val_loss: 0.0251\n",
+ "Epoch 17/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m18s\u001b[0m 253ms/step - accuracy: 0.9772 - loss: 0.0755 - val_accuracy: 0.9839 - val_loss: 0.0446\n",
+ "Epoch 18/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m23s\u001b[0m 290ms/step - accuracy: 0.9908 - loss: 0.0262 - val_accuracy: 0.9875 - val_loss: 0.0290\n",
+ "Epoch 19/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m19s\u001b[0m 265ms/step - accuracy: 0.9895 - loss: 0.0328 - val_accuracy: 0.9848 - val_loss: 0.0459\n",
+ "Epoch 20/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 263ms/step - accuracy: 0.9862 - loss: 0.0358 - val_accuracy: 0.9875 - val_loss: 0.0359\n",
+ "Epoch 21/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 233ms/step - accuracy: 0.9959 - loss: 0.0135 - val_accuracy: 0.9964 - val_loss: 0.0081\n",
+ "Epoch 22/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 229ms/step - accuracy: 0.9968 - loss: 0.0077 - val_accuracy: 0.9946 - val_loss: 0.0132\n",
+ "Epoch 23/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m23s\u001b[0m 260ms/step - accuracy: 0.9919 - loss: 0.0270 - val_accuracy: 0.9875 - val_loss: 0.0398\n",
+ "Epoch 24/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 264ms/step - accuracy: 0.9944 - loss: 0.0231 - val_accuracy: 0.9982 - val_loss: 0.0041\n",
+ "Epoch 25/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 232ms/step - accuracy: 0.9919 - loss: 0.0241 - val_accuracy: 0.9661 - val_loss: 0.0930\n",
+ "Epoch 26/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 246ms/step - accuracy: 0.9942 - loss: 0.0213 - val_accuracy: 0.9402 - val_loss: 0.2238\n",
+ "Epoch 27/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 217ms/step - accuracy: 0.9879 - loss: 0.0354 - val_accuracy: 0.9946 - val_loss: 0.0154\n",
+ "Epoch 28/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 224ms/step - accuracy: 0.9924 - loss: 0.0232 - val_accuracy: 0.9946 - val_loss: 0.0179\n",
+ "Epoch 29/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 214ms/step - accuracy: 0.9978 - loss: 0.0098 - val_accuracy: 0.9982 - val_loss: 0.0081\n",
+ "Epoch 30/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m23s\u001b[0m 252ms/step - accuracy: 0.9981 - loss: 0.0047 - val_accuracy: 1.0000 - val_loss: 0.0010\n",
+ "Epoch 31/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 218ms/step - accuracy: 0.9927 - loss: 0.0223 - val_accuracy: 0.9911 - val_loss: 0.0285\n",
+ "Epoch 32/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 234ms/step - accuracy: 0.9906 - loss: 0.0347 - val_accuracy: 0.9955 - val_loss: 0.0136\n",
+ "Epoch 33/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m22s\u001b[0m 263ms/step - accuracy: 0.9988 - loss: 0.0070 - val_accuracy: 0.9982 - val_loss: 0.0059\n",
+ "Epoch 34/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 232ms/step - accuracy: 0.9978 - loss: 0.0091 - val_accuracy: 0.9991 - val_loss: 0.0019\n",
+ "Epoch 35/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m23s\u001b[0m 263ms/step - accuracy: 0.9993 - loss: 0.0036 - val_accuracy: 0.9964 - val_loss: 0.0123\n",
+ "Epoch 36/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 265ms/step - accuracy: 0.9869 - loss: 0.0485 - val_accuracy: 0.9991 - val_loss: 0.0031\n",
+ "Epoch 37/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 279ms/step - accuracy: 0.9973 - loss: 0.0094 - val_accuracy: 0.9866 - val_loss: 0.0475\n",
+ "Epoch 38/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m19s\u001b[0m 251ms/step - accuracy: 0.9921 - loss: 0.0277 - val_accuracy: 0.9875 - val_loss: 0.0415\n",
+ "Epoch 39/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 251ms/step - accuracy: 0.9939 - loss: 0.0261 - val_accuracy: 1.0000 - val_loss: 0.0015\n",
+ "Epoch 40/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 252ms/step - accuracy: 0.9998 - loss: 0.0023 - val_accuracy: 1.0000 - val_loss: 6.3795e-04\n",
+ "Epoch 41/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 253ms/step - accuracy: 0.9996 - loss: 0.0017 - val_accuracy: 1.0000 - val_loss: 4.4335e-04\n",
+ "Epoch 42/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 227ms/step - accuracy: 1.0000 - loss: 7.3446e-04 - val_accuracy: 1.0000 - val_loss: 1.9014e-04\n",
+ "Epoch 43/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m22s\u001b[0m 254ms/step - accuracy: 1.0000 - loss: 9.8123e-04 - val_accuracy: 1.0000 - val_loss: 1.3924e-04\n",
+ "Epoch 44/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 230ms/step - accuracy: 0.9986 - loss: 0.0042 - val_accuracy: 0.9911 - val_loss: 0.0443\n",
+ "Epoch 45/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m18s\u001b[0m 255ms/step - accuracy: 0.9917 - loss: 0.0245 - val_accuracy: 0.9955 - val_loss: 0.0122\n",
+ "Epoch 46/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 257ms/step - accuracy: 0.9925 - loss: 0.0294 - val_accuracy: 0.9929 - val_loss: 0.0243\n",
+ "Epoch 47/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 248ms/step - accuracy: 0.9958 - loss: 0.0103 - val_accuracy: 1.0000 - val_loss: 0.0021\n",
+ "Epoch 48/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m22s\u001b[0m 270ms/step - accuracy: 0.9993 - loss: 0.0027 - val_accuracy: 1.0000 - val_loss: 3.4261e-04\n",
+ "Epoch 49/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 245ms/step - accuracy: 0.9985 - loss: 0.0039 - val_accuracy: 0.9964 - val_loss: 0.0065\n",
+ "Epoch 50/50\n",
+ "\u001b[1m70/70\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 248ms/step - accuracy: 0.9989 - loss: 0.0020 - val_accuracy: 1.0000 - val_loss: 2.4008e-04\n"
+ ]
+ }
+ ],
+ "source": [
+ "# Train the model\n",
+ "history = model.fit(X, y, validation_split=0.2, epochs=50, batch_size=64)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# best val accuracy: 72.32\n",
+ "# use checkpoint to save the best val accuracy model\n",
+ "# adjust learning rate for slow convergence"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 29,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "epochs = list(range(50))\n",
+ "acc = history.history['accuracy']\n",
+ "val_acc = history.history['val_accuracy']\n",
+ "\n",
+ "plt.plot(epochs, acc, label='train accuracy')\n",
+ "plt.plot(epochs, val_acc, label='val accuracy')\n",
+ "plt.xlabel('epochs')\n",
+ "plt.ylabel('accuracy')\n",
+ "plt.legend()\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 30,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "loss = history.history['loss']\n",
+ "val_loss = history.history['val_loss']\n",
+ "\n",
+ "plt.plot(epochs, loss, label='train loss')\n",
+ "plt.plot(epochs, val_loss, label='val loss')\n",
+ "plt.xlabel('epochs')\n",
+ "plt.ylabel('loss')\n",
+ "plt.legend()\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "base",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.11.5"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git a/Emotion_Recognition_Audio/requirements.txt b/Emotion_Recognition_Audio/requirements.txt
new file mode 100644
index 000000000..1c255c2f5
--- /dev/null
+++ b/Emotion_Recognition_Audio/requirements.txt
@@ -0,0 +1,8 @@
+pandas==1.5.3
+numpy==1.23.5
+seaborn==0.12.2
+matplotlib==3.7.1
+librosa==0.10.0
+IPython==8.13.2
+scikit-learn==1.2.2
+tensorflow==2.12.0
\ No newline at end of file
From cb11b23741d316d4abca3007a47c162348ca365a Mon Sep 17 00:00:00 2001
From: Chethana Potukanam
<142076224+ChethanaPotukanam@users.noreply.github.com>
Date: Wed, 17 Jul 2024 23:02:04 +0530
Subject: [PATCH 11/14] Delete Emotion_Recognition_Audio/README.md
---
Emotion_Recognition_Audio/README.md | 1 -
1 file changed, 1 deletion(-)
delete mode 100644 Emotion_Recognition_Audio/README.md
diff --git a/Emotion_Recognition_Audio/README.md b/Emotion_Recognition_Audio/README.md
deleted file mode 100644
index 8b1378917..000000000
--- a/Emotion_Recognition_Audio/README.md
+++ /dev/null
@@ -1 +0,0 @@
-
From c6d69e61199127aaab4e893a49a48be45591caba Mon Sep 17 00:00:00 2001
From: Chethana Potukanam
<142076224+ChethanaPotukanam@users.noreply.github.com>
Date: Wed, 17 Jul 2024 23:02:26 +0530
Subject: [PATCH 12/14] Add files via upload
---
Emotion_Recognition_Audio/README.md | 26 ++++++++++++++++++++++++++
1 file changed, 26 insertions(+)
create mode 100644 Emotion_Recognition_Audio/README.md
diff --git a/Emotion_Recognition_Audio/README.md b/Emotion_Recognition_Audio/README.md
new file mode 100644
index 000000000..c02c809a3
--- /dev/null
+++ b/Emotion_Recognition_Audio/README.md
@@ -0,0 +1,26 @@
+# Emotion Recognition from Audio using Deep Learning
+
+This project aims to recognize emotions from audio files using deep learning. The pipeline includes loading and processing audio data, visualizing waveforms and spectrograms, extracting MFCC features, and training a neural network model.
+
+Dataset Download link: https://www.kaggle.com/ejlok1/toronto-emotional-speech-set-tess More Datasets: https://www.kaggle.com/dmitrybabko/speech-emotion-recognition-en
+
+## Project Structure
+
+my_project/
+│
+├── dataset/
+│
+├── main.ipynb
+├── README.md
+└── requirements.txt
+
+
+## Installation
+
+1. Clone the repository
+ ```sh
+ git clone https://github.com/ChethanaPotukanam/EmotionRecognition.git
+ cd EmotionRecognition
+
+#### To install the required packages
+pip install -r requirements.txt
\ No newline at end of file
From 853348ff23c3446c305a092dc32c237c7979146b Mon Sep 17 00:00:00 2001
From: Chethana Potukanam
<142076224+ChethanaPotukanam@users.noreply.github.com>
Date: Wed, 17 Jul 2024 23:03:56 +0530
Subject: [PATCH 13/14] Delete Emotion_Recognition_Audio/README.md
---
Emotion_Recognition_Audio/README.md | 26 --------------------------
1 file changed, 26 deletions(-)
delete mode 100644 Emotion_Recognition_Audio/README.md
diff --git a/Emotion_Recognition_Audio/README.md b/Emotion_Recognition_Audio/README.md
deleted file mode 100644
index c02c809a3..000000000
--- a/Emotion_Recognition_Audio/README.md
+++ /dev/null
@@ -1,26 +0,0 @@
-# Emotion Recognition from Audio using Deep Learning
-
-This project aims to recognize emotions from audio files using deep learning. The pipeline includes loading and processing audio data, visualizing waveforms and spectrograms, extracting MFCC features, and training a neural network model.
-
-Dataset Download link: https://www.kaggle.com/ejlok1/toronto-emotional-speech-set-tess More Datasets: https://www.kaggle.com/dmitrybabko/speech-emotion-recognition-en
-
-## Project Structure
-
-my_project/
-│
-├── dataset/
-│
-├── main.ipynb
-├── README.md
-└── requirements.txt
-
-
-## Installation
-
-1. Clone the repository
- ```sh
- git clone https://github.com/ChethanaPotukanam/EmotionRecognition.git
- cd EmotionRecognition
-
-#### To install the required packages
-pip install -r requirements.txt
\ No newline at end of file
From a090d126be4bc8233b3fcee06006b7ea985711c3 Mon Sep 17 00:00:00 2001
From: Chethana Potukanam
<142076224+ChethanaPotukanam@users.noreply.github.com>
Date: Wed, 17 Jul 2024 23:04:30 +0530
Subject: [PATCH 14/14] Add files via upload
---
Emotion_Recognition_Audio/README.md | 25 +++++++++++++++++++++++++
1 file changed, 25 insertions(+)
create mode 100644 Emotion_Recognition_Audio/README.md
diff --git a/Emotion_Recognition_Audio/README.md b/Emotion_Recognition_Audio/README.md
new file mode 100644
index 000000000..f2d3df893
--- /dev/null
+++ b/Emotion_Recognition_Audio/README.md
@@ -0,0 +1,25 @@
+# Emotion Recognition from Audio using Deep Learning
+
+This project aims to recognize emotions from audio files using deep learning. The pipeline includes loading and processing audio data, visualizing waveforms and spectrograms, extracting MFCC features, and training a neural network model.
+
+Dataset Download link: https://www.kaggle.com/ejlok1/toronto-emotional-speech-set-tess More Datasets: https://www.kaggle.com/dmitrybabko/speech-emotion-recognition-en
+
+## Project Structure
+
+my_project/
+│
+├── dataset/
+│
+├── main.ipynb
+├── README.md
+└── requirements.txt
+
+
+## Installation
+
+1. Clone the repository
+ ```sh
+ https://github.com/ChethanaPotukanam/DL-Simplified/tree/main/Emotion_Recognition_Audio
+
+#### To install the required packages
+pip install -r requirements.txt
\ No newline at end of file